CN114168216B - Parameter tuning method, device and storage medium - Google Patents
Parameter tuning method, device and storage medium Download PDFInfo
- Publication number
- CN114168216B CN114168216B CN202111404891.3A CN202111404891A CN114168216B CN 114168216 B CN114168216 B CN 114168216B CN 202111404891 A CN202111404891 A CN 202111404891A CN 114168216 B CN114168216 B CN 114168216B
- Authority
- CN
- China
- Prior art keywords
- tuning
- parameters
- tuning effect
- parameter
- test
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 99
- 238000003860 storage Methods 0.000 title claims abstract description 15
- 230000000694 effects Effects 0.000 claims abstract description 209
- 238000012360 testing method Methods 0.000 claims abstract description 163
- 230000008569 process Effects 0.000 claims abstract description 64
- 238000004088 simulation Methods 0.000 claims abstract description 62
- 238000004458 analytical method Methods 0.000 claims description 36
- 238000010801 machine learning Methods 0.000 claims description 36
- 230000035945 sensitivity Effects 0.000 claims description 15
- 238000005516 engineering process Methods 0.000 claims description 13
- 238000013139 quantization Methods 0.000 abstract description 6
- 238000010586 diagram Methods 0.000 description 12
- 238000004891 communication Methods 0.000 description 10
- 238000004590 computer program Methods 0.000 description 10
- 238000012545 processing Methods 0.000 description 7
- 239000011159 matrix material Substances 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 230000002776 aggregation Effects 0.000 description 3
- 238000004220 aggregation Methods 0.000 description 3
- 230000004927 fusion Effects 0.000 description 3
- 238000012417 linear regression Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000006116 polymerization reaction Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/445—Program loading or initiating
- G06F9/44568—Immediately runnable code
- G06F9/44578—Preparing or optimising for loading
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Preventing errors by testing or debugging software
- G06F11/3668—Software testing
- G06F11/3672—Test management
- G06F11/3688—Test management for test execution, e.g. scheduling of test suites
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/217—Database tuning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/30—Creation or generation of source code
- G06F8/37—Compiler construction; Parser generation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/4401—Bootstrapping
- G06F9/4406—Loading of operating system
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Quality & Reliability (AREA)
- Computer Hardware Design (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Computer Security & Cryptography (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the application provides a parameter tuning method, equipment and a storage medium. In the embodiment of the application, the target parameter set to be tuned can be subjected to a simulated tuning test to obtain a tuning effect test value; according to the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values, determining the nonlinear correlation degree between each of the plurality of parameters and the tuning effect; selecting sensitive parameters from the target parameter group according to the nonlinear correlation degree; and performing parameter tuning on the object to be tuned according to the sensitive parameters. Accordingly, in the embodiment of the application, the nonlinear relation between the quantization parameter and the tuning effect is provided in the parameter tuning process, so that the complex relation between the parameter and the tuning effect is better embodied, the sensitive parameter is more accurately screened out, the tuning search space is reduced based on the sensitive parameter, and the tuning efficiency is further effectively improved.
Description
Technical Field
The present application relates to the field of data processing technologies, and in particular, to a parameter tuning method, apparatus, and storage medium.
Background
The tuning requirements for parameters exist in the fields of an operating system kernel, a compiler, a database and the like, and the service running performance can be optimized by relatively effective parameter setting, so that the overall service quality is improved.
The current mainstream tuning algorithm is based on bayesian theory. And according to the tuning effect fed back in real time, carrying out dynamic search in a preset parameter space. However, the number of parameters involved in the above field is large, possibly up to hundreds of dimensions, and the search period required when the bayesian algorithm is applied will be very long, which not only results in low tuning efficiency, but also may affect the tuning effect.
Disclosure of Invention
The application provides a parameter tuning method and equipment, which are used for improving parameter tuning efficiency.
The embodiment of the application provides a parameter tuning method, which comprises the following steps:
obtaining a target parameter set to be optimized, wherein the target parameter set comprises a plurality of parameters;
performing simulation tuning test on the target parameter set to obtain a tuning effect test value;
according to the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values, determining the nonlinear correlation degree between each of the plurality of parameters and the tuning effect;
selecting sensitive parameters from the target parameter group according to the nonlinear correlation degree;
and performing parameter tuning on the object to be tuned according to the sensitive parameters.
The embodiment of the application also provides a computing device, which comprises a memory and a processor;
The memory is used for storing one or more computer instructions;
the processor is coupled to the memory for executing the one or more computer instructions for:
obtaining a target parameter set to be optimized, wherein the target parameter set comprises a plurality of parameters;
performing simulation tuning test on the target parameter set to obtain a tuning effect test value;
according to the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values, determining the nonlinear correlation degree between each of the plurality of parameters and the tuning effect;
selecting sensitive parameters from the target parameter group according to the nonlinear correlation degree;
and performing parameter tuning on the object to be tuned according to the sensitive parameters.
Embodiments of the present application also provide a computer-readable storage medium storing computer instructions that, when executed by one or more processors, cause the one or more processors to perform the foregoing parameter tuning method.
In the embodiment of the application, the target parameter set to be tuned can be subjected to a simulated tuning test to obtain a tuning effect test value; according to the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values, determining the nonlinear correlation degree between each of the plurality of parameters and the tuning effect; selecting sensitive parameters from the target parameter group according to the nonlinear correlation degree; and performing parameter tuning on the object to be tuned according to the sensitive parameters. Accordingly, in the embodiment of the application, the nonlinear relation between the quantization parameter and the tuning effect is provided in the parameter tuning process, so that the complex relation between the parameter and the tuning effect is better embodied, the sensitive parameter is more accurately screened out, the tuning search space is reduced based on the sensitive parameter, and the tuning efficiency is further effectively improved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute a limitation on the application. In the drawings:
FIG. 1 is a schematic flow chart of a parameter tuning method according to an exemplary embodiment of the present application;
Fig. 2 is a schematic structural diagram of a parameter tuning device according to an exemplary embodiment of the present application;
FIG. 3 is a schematic diagram of a nonlinear analyzer according to an exemplary embodiment of the present application;
Fig. 4 is a schematic structural diagram of a computing device according to another exemplary embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be clearly and completely described below with reference to specific embodiments of the present application and corresponding drawings. It will be apparent that the described embodiments are only some, but not all, embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
At present, the main stream tuning algorithm is based on the Bayesian theory, but under the condition of a large number of parameters, the search period required by applying the Bayesian algorithm is very long, which not only results in low tuning efficiency, but also can influence the tuning effect. To this end, in some embodiments of the application: the target parameter set to be tuned can be subjected to a simulated tuning test to obtain a tuning effect test value; according to the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values, determining the nonlinear correlation degree between each of the plurality of parameters and the tuning effect; selecting sensitive parameters from the target parameter group according to the nonlinear correlation degree; and performing parameter tuning on the object to be tuned according to the sensitive parameters. Accordingly, in the embodiment of the application, the nonlinear relation between the quantization parameter and the tuning effect is provided in the parameter tuning process, so that the complex relation between the parameter and the tuning effect is better embodied, the sensitive parameter is more accurately screened out, the tuning search space is reduced based on the sensitive parameter, and the tuning efficiency is further effectively improved.
The following describes in detail the technical solutions provided by the embodiments of the present application with reference to the accompanying drawings.
Fig. 1 is a flow chart of a parameter tuning method according to an exemplary embodiment of the present application, where the method may be performed by a parameter tuning device, and the parameter tuning device may be implemented as a combination of software and/or hardware, and the parameter tuning device may be integrated in a computing device. Referring to fig. 1, the method includes:
step 100, obtaining a target parameter set to be optimized, wherein the target parameter set comprises a plurality of parameters;
Step 101, performing simulation tuning test on a target parameter set to obtain a tuning effect test value;
step 102, determining the nonlinear correlation degree between each of a plurality of parameters and the tuning effect according to the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values;
step 103, selecting sensitive parameters from the target parameter group according to the nonlinear relativity;
And 104, performing parameter tuning according to the sensitive parameters.
The parameter tuning method provided by the embodiment can be applied to various scenes with parameter tuning requirements. For example, the operating system kernel tuning, the compiler tuning, the database tuning, and the like, and the application scenario is not particularly limited in this embodiment. In different application scenarios, there may be a difference between the target parameter sets to be tuned, and in this embodiment, the target parameter sets may be determined according to actual requirements. In this embodiment, the number, type, etc. of parameters in the target parameter set are not limited. For example, where network performance is a concern, core parameters related to network performance may be mapped into a set of target parameters. In this way, in the present embodiment, one or more parameters may be included in the target parameter set, and for the case that the target parameter set includes only one parameter, the parameter may be directly tuned, but in the present embodiment, focus is focused on the case that the target parameter set includes a plurality of parameters, so as to improve the parameter tuning efficiency in this case.
In step 101, a simulation tuning test may be performed on the target parameter set to obtain a tuning effect test value. The tuning effect in this embodiment may include one or more dimensions, or take the above case of focusing on network performance as an example, where the tuning effect may include multiple dimensions such as throughput and delay. In this embodiment, multiple simulation tuning tests may be performed on the target parameter set, and each simulation tuning test may generate tuning effect test values of each dimension. In an alternative implementation: and a benchmark test benchmark technology can be adopted to perform simulation tuning test on the target parameter set so as to obtain a tuning effect test value. In this exemplary scheme, in the single simulation tuning test process, the parameter values in the target parameter set may be adjusted, and the tuning effect of the real scene may be simulated by using benchemark technology according to these parameter values, so as to generate tuning effect test values in one or more dimensions. Fig. 2 is a schematic structural diagram of a parameter tuning device according to an exemplary embodiment of the present application, and referring to fig. 2, the parameter tuning device includes an input module 10, where the input module 10 may be used to perform the operation of step 101, and provide the processing result to a relationship identifying module 20.
Through the simulation tuning test process, a data set [ x1, y 1), (x 2, y 2) … (xi, yi) … (xn, yn) ] can be obtained for the tuning effect of a single dimension, wherein n represents the simulation tuning test times, xi represents the parameter value in the target parameter set in the ith simulation tuning test process, and yi represents the tuning effect test value obtained in the ith simulation tuning test process. Under the condition that the tuning effect comprises a plurality of dimensions, corresponding data sets can be obtained under the tuning effect of other dimensions. It should be noted that the above data set is only used to illustrate the parameter values of the target parameter set in the process of the analog tuning test and the states of the corresponding tuning effect test values, and is not limited to the recording form of these data. Optionally, the data set may be recorded in a matrix form, for example, the target parameter set includes 100 parameters, and the parameter values in the 10-time simulation tuning test process may be recorded in a 100 x 10-dimensional matrix, where each row in the matrix is used to record 100 parameter values in the single simulation tuning test process; the 10 tuning effect test values can be recorded in a1 x 10-dimensional matrix, and each row in the matrix is used for recording the tuning effect test values in the single simulation tuning test process. Of course, this is merely an example, and the present embodiment is not limited thereto. In addition, in the present embodiment, the larger the data amount in the above-described data set, the higher the processing accuracy in the subsequent step will be, as conditions permit.
Based on this, in step 102, the nonlinear correlation between each of the plurality of parameters and the tuning effect is determined by using the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values as analysis basis. Specifically, in the case where the tuning effect includes multiple dimensions, in step 102, a nonlinear correlation between each of the multiple parameters and the tuning effect of a certain dimension may be determined, for example, a nonlinear correlation between each of the multiple parameters and the network throughput may be determined, and a nonlinear correlation between each of the multiple parameters and the network delay may also be determined. In this embodiment, the nonlinear correlation may be used to characterize a nonlinear relationship between the parameter and the tuning effect. The generation scheme regarding the nonlinear correlation will be described in detail later.
In this embodiment, it is proposed to quantitatively analyze a nonlinear relationship between the parameter and the tuning effect, and the influence relationship between the parameter and the tuning effect can be more accurately and more perfectly represented based on the nonlinear correlation. Referring to fig. 2, the relationship identification module 20 in the parameter tuning device may be used to perform the operation in step 102, wherein the relationship identification module 20 may include a nonlinear analyzer 21 therein, and accordingly, the nonlinear analyzer 21 in the relationship identification module 20 may perform the related operation of step 102.
Based on this, in step 103, a sensitive parameter may be selected from the set of target parameters according to the non-linear correlation. The influence relation between each of the plurality of parameters in the target parameter set and the tuning effect can be more accurately represented based on the nonlinear correlation, so that the sensitivity degree of the tuning effect to the tuning action of the plurality of parameters can be more accurately found, and the sensitive parameters corresponding to the tuning effect can be selected from the target parameter set. In this embodiment, the sensitive parameter refers to a parameter that is more sensitive to the influence of the tuning effect when the tuning operation occurs thereon. It will be appreciated that the sensitive parameters are typically part of the set of target parameters, so that in step 103, the optimized search space may be reduced to include only sensitive parameters, which are much smaller than the original set of target parameters, and thus, optimization of the optimized search space may be achieved. And as the nonlinear relativity is introduced to participate in the influence relation between the characterization parameters and the tuning effect, the sensitive parameters can be very accurately selected, and the tuning effect can be effectively ensured. Referring to fig. 2, the output module 30 in the parameter tuning device may be used to perform the steps of step 103.
Of course, in this embodiment, the influence relationship between the parameter and the tuning effect is not limited to be represented by only relying on the nonlinear correlation, but the representation effect of the influence relationship between the parameter and the tuning effect is added by the nonlinear correlation. In this embodiment, besides the nonlinear correlation, the mutual information values and/or the linear correlation between the multiple parameters in the target parameter set and the tuning effect may be obtained to cooperatively represent the influence relationship between the multiple parameters and the tuning effect.
In step 103, a linear correlation and/or a mutual information value between each of the plurality of parameters and the tuning effect may be obtained; correcting the nonlinear correlation according to the linear correlation and/or the mutual information value to obtain the sensitivity indexes corresponding to the parameters; and selecting the sensitive parameters from the target parameter group according to the sensitive index. The mutual information can be used for representing mutual information between each of the plurality of parameters and the tuning effect, and can represent whether the relationship exists between the parameters and the tuning effect, the strength of the relationship and the like. The linear correlation may be used to characterize the linear relationship between each of the plurality of parameters and the tuning effect. The generation scheme regarding the mutual information value and the linear correlation will be described in detail later. In this way, in this embodiment, by correcting the nonlinear correlation by the mutual information value and the linear correlation, the influence relationship between the parameter and the tuning effect can be more accurately represented, so as to ensure the accuracy and stability of the identification of the sensitivity coefficient. Referring to fig. 2, the relationship identifying module 20 in the parameter tuning device may further include a mutual information analyzer 22, a linear analyzer 23 and a fusion device 24, where the mutual information analyzer 22 may be configured to obtain mutual information values between each of the plurality of parameters and the tuning effect, the linear analyzer may be configured to obtain linear correlations between each of the plurality of parameters and the tuning effect, and the fusion device 24 may be configured to correct the nonlinear correlations according to the linear correlations and/or the mutual information values, so as to obtain sensitivity indexes corresponding to each of the plurality of parameters.
After the sensitive parameters are determined, in this embodiment, parameter tuning may be performed according to the sensitive parameters. Even if tuning operation can be performed only for sensitive parameters, and tuning operation is not performed on other parameters in the target parameter set, the tuning search space can be greatly reduced, and therefore parameter tuning efficiency is effectively improved. Referring to fig. 2, the parameter tuning device may further include a tuning module 40 for tuning parameters according to sensitive parameters.
In this embodiment, parameter self-tuning may be implemented according to the sensitive parameter, and an exemplary self-tuning scheme may be: in a given parameter value space, according to a certain algorithm, the proper parameter value is automatically searched and solved in the parameter value space according to real-time feedback. Thus, the intelligent and automatic parameter tuning can be realized in the embodiment. The specific tuning logic for parameter tuning is not limited in this embodiment.
Accordingly, in this embodiment, the target parameter set to be tuned may be subjected to a simulated tuning test to obtain a tuning effect test value; according to the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values, determining the nonlinear correlation degree between each of the plurality of parameters and the tuning effect; selecting sensitive parameters from the target parameter group according to the nonlinear correlation degree; and performing parameter tuning on the object to be tuned according to the sensitive parameters. Accordingly, in the embodiment of the application, the nonlinear relation between the quantization parameter and the tuning effect is provided in the parameter tuning process, so that the complex relation between the parameter and the tuning effect is better embodied, the sensitive parameter is more accurately screened out, the tuning search space is reduced based on the sensitive parameter, and the tuning efficiency is further effectively improved.
In the above or below embodiments, various implementations may be used to generate a nonlinear correlation between each of the multiple parameters in the target parameter set and the tuning effect.
In an alternative implementation manner, the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values can be provided for the machine learning model, so that the machine learning model captures the nonlinear relation between each of a plurality of parameters in the target parameter set and the tuning effect; and quantifying the nonlinear relation captured by the machine learning model by using a model interpreter to generate nonlinear correlation between each of the plurality of parameters and the tuning effect. The machine learning model may be a nonlinear machine learning model, including but not limited to a polynomial model, a support vector machine model, a decision tree model, various neural network models, etc., and may be selected according to the needs in practical applications.
In the implementation manner, the machine learning model can be trained by taking the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values as training samples, so that the machine learning model captures the nonlinear relation between each of a plurality of parameters in the target parameter set and the tuning effect. The complexity of the nonlinear machine learning model is high, so that in the implementation mode, the machine learning model can be treated as a black box, and the machine learning model is interpreted by adopting an interpretable artificial intelligence technology, so that the nonlinear relation captured by the machine learning model is quantized.
Fig. 3 is a schematic structural diagram of a nonlinear analyzer according to an exemplary embodiment of the present application, and referring to fig. 3, a nonlinear machine learning model 210 and a model interpreter 211 may be included in the nonlinear analyzer 21. The nonlinear machine learning model may learn a nonlinear relationship between the parameters and the tuning effect, and for this purpose, the aforementioned data sets [ X1, Y1), (X2, Y2) … (xi, yi) … (xn, yn) ] may be input as training samples to the nonlinear machine learning model 210 for learning the nonlinear relationship between X (as model input data) and Y (as model prediction result). For the model interpreter 211, the recognition degree of the correlation between the input data and the predicted result by the nonlinear machine learning model 210 can be estimated through a statistical or game theory method, so as to quantify the nonlinear relation captured by the nonlinear machine learning model 210.
In this implementation, the quantization process may be: the machine learning model can be utilized to construct more parameter values and corresponding tuning effect test values in the simulation tuning test process according to the captured nonlinear relation; providing the parameter values generated by performing simulation tuning test operation on the target parameter set and constructed by the machine learning model and the corresponding tuning effect test values to a model interpreter; and generating nonlinear correlations between each of the plurality of parameters and the tuning effect by using a model interpreter. Referring to fig. 2, the nonlinear machine learning model 210 may construct more parameter values and corresponding tuning effect test values in the simulation tuning test process according to the captured nonlinear relationship, so as to enhance the data set generated by the simulation tuning test operation in the foregoing step 101. For example, n in the dataset may be enhanced from 100 to 1000. Thus, with sufficient data support, the model interpreter 211 may quantify the non-linear relationship between each of the plurality of parameters in the set of target parameters and the tuning effect by means of statistical or game theory methods.
In one exemplary scenario: in a model interpreter, adopting a saprolite shape algorithm, and analyzing the contribution degree of each of a plurality of parameters to the tuning effect according to the parameter value and the corresponding tuning effect test value; and determining the nonlinear correlation between each of the plurality of parameters and the tuning effect based on the contribution degree of each of the plurality of parameters to the tuning effect. The saprolitic algorithm may explain the contribution degree of different input data to the prediction result in the machine learning model, and in this exemplary scheme, the absolute value of the contribution degree of each of the plurality of parameters to the tuning effect may be normalized, so as to obtain a nonlinear correlation degree between each of the plurality of parameters and the tuning effect. For example, if n in the data set generated by the analog tuning test operation is enhanced from 100 to 1000, which is equivalent to providing 1000 times of the analog tuning test process (the parameter value, the tuning effect test value) to the model interpreter, the model interpreter may output contribution degrees corresponding to the parameters for each time of the analog tuning test process (the parameter value, the tuning effect test value), for example, 200 parameters are included in the target parameter set, and the model interpreter may output 200 contribution degrees for one time of the analog tuning test process. After 1000 times of quantization, each parameter will obtain 1000 contribution degrees for the tuning effect, where a final contribution degree can be generated by averaging, median, etc. of the 1000 contribution degrees of the parameter a, and the final contribution degree is used as the contribution degree of the parameter a for the tuning effect, and similarly, the contribution degrees of other 199 parameters for the tuning effect can be obtained. The contribution of each of the 200 parameters to the tuning effect may then be normalized to produce a nonlinear correlation between each of the 200 parameters and the tuning effect.
It should be understood that other implementations may be used to generate the nonlinear correlation between each of the plurality of parameters in the target parameter set and the tuning effect in the present embodiment, and the present embodiment is not limited thereto, for example, a sufficient number of analog tuning tests may be performed in step 101 to generate a sufficient number of data sets, and statistics or game theory may be applied on the basis of the sufficient number of data sets to quantify the nonlinear correlation between each of the plurality of parameters in the target parameter set and the tuning effect.
In this embodiment, various implementations may be used to generate the mutual information values between the tuning effects and the respective parameters in the target parameter set.
In an alternative implementation manner, mutual information analysis can be performed on the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values so as to generate mutual information values between each of the plurality of parameters and the tuning effect. In this implementation, the mutual information analyzer 22 of fig. 2 may construct a mutual information analysis model, which may employ a machine learning model, and may capture the correlation degree of xi and yi in the data set generated by the aforementioned analog tuning test procedure as the mutual information value based mainly on the principle of mutual information (Mutual Information). This principle assumes that multiple parameters in the target parameter set are independent of each other and thus only concerns the statistical correlation between a single parameter and the tuning effect. Therefore, the correlation between the multiple parameters in the target parameter set and the tuning effect can be comprehensively captured through mutual information analysis, wherein the correlation comprises linear correlation and nonlinear correlation. On the basis, the absolute value of the mutual information analysis result between each of the plurality of parameters and the tuning parameter can be normalized to a 0-1 interval and then used as the mutual information value between each of the plurality of parameters and the tuning parameter.
Taking the target tuning effect as an example when the tuning effect comprises multiple dimensions, selecting at least one tuning effect test value corresponding to the target tuning effect from tuning effect test values generated in the simulation tuning test process and taking the parameter value in the at least one target simulation tuning test process; performing mutual information analysis on the parameter values and the corresponding tuning effect test values in the at least one target simulation test process to generate mutual information values between a plurality of parameters and the target tuning effect; the target tuning effect is any one of tuning effects. For example, if the number of analog tuning tests for the network throughput is 100, the (parameter value, tuning effect test value) in the 100 analog tuning test processes may be obtained, 100 sets (parameter value, tuning effect test value) may be provided to the mutual information analyzer in fig. 2, if the target parameter set includes 200 parameters, the mutual information analyzer may output 200 mutual information analysis results for each set (parameter value, tuning effect test value), so that 100 mutual information analysis results between the parameter a and the network throughput may be obtained, where the 100 mutual information analysis results of the parameter a may be averaged, median value may be obtained, and so on to generate a final mutual information analysis result as a final mutual information analysis result between the parameter a and the network throughput, and similarly, final mutual information analysis results between other 199 parameters in the target parameter set and the network throughput may be obtained. The absolute values of the mutual information analysis results between the 200 parameters and the network throughput may then be normalized to produce mutual information values between the 200 parameters and the network throughput.
It should be understood that other implementations may be used in the present embodiment to generate the mutual information values between each of the plurality of parameters in the target parameter set and the tuning effect, and the present embodiment is not limited thereto.
In this embodiment, various implementations may be used to generate the linear correlation between each of the multiple parameters in the target parameter set and the tuning effect.
In an alternative implementation manner, the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values can be linearly analyzed to generate the linear correlation degree between each of the plurality of parameters and the tuning effect. In this implementation manner, the linear analyzer 23 in fig. 2 may construct a linear regression model f_l for learning a functional relationship between the parameter value and the tuning effect test value according to the parameter value and the corresponding tuning effect test value of the target parameter set in the simulation tuning test process, that is, letting the model prediction value be f_l (X; W) =sum { i=1,., n } (w_i×i+b_i), where i represents the i-th simulation tuning test, xi represents the parameter value in the i-th simulation tuning test process, and wi represents the linear coefficient corresponding to each of the plurality of parameters in the i-th simulation tuning test process. The model parameters w= \ { w_i, w_2, & gt, W n\ }, the absolute values of the model parameters W are normalized to the interval of 0-1 and can be used as the linear correlation degree based on linear regression identification.
Taking the target tuning effect as an example when the tuning effect comprises multiple dimensions, selecting at least one tuning effect test value corresponding to the target tuning effect from tuning effect test values generated in the simulation tuning test process and taking the parameter value in the at least one target simulation tuning test process; performing linear analysis on the parameter values and the corresponding tuning effect test values in the at least one target simulation test process to generate linear correlation between a plurality of parameters and the target tuning effect; the target tuning effect is any one of tuning effects. For example, if the number of analog tuning tests for the network throughput is 100, the (parameter value, tuning effect test value) in the 100 analog tuning test processes may be obtained, 100 sets (parameter value, tuning effect test value) are provided to the linear analyzer in fig. 2, if the target parameter set includes 200 parameters, the linear analyzer may output 200 linear analysis results for each set (parameter value, tuning effect test value), so 100 linear analysis results between the parameter a and the network throughput may be obtained, where the 100 linear analysis results of the parameter a may be averaged, median value, and the like to generate a final linear analysis result as a final linear analysis result between the parameter a and the network throughput, and similarly, final linear analysis results between other 199 parameters in the target parameter set and the network throughput may be obtained. The absolute values of the linear analysis results between the 200 parameters and the network throughput may then be normalized to produce a linear correlation between the 200 parameters and the network throughput.
Referring to fig. 2, the fusion device 24 in the relationship identification module 20 may be used to aggregate the analysis results output by the mutual information analyzer, the linear analyzer, and the nonlinear analyzer. An exemplary manner of aggregation may be: determining weights corresponding to the linear correlation, the mutual information value and the nonlinear correlation; under the target parameters, weighting linear correlation, mutual information value and nonlinear correlation according to the weights to obtain sensitivity coefficients corresponding to the target parameters; wherein the target parameter is any one of a plurality of parameters. Here, a linear weighted aggregation mode is adopted, and each weight can be adjusted according to actual needs. In terms of parameter tuning, since the analysis results of the mutual information analyzer and the linear analyzer are mainly used as a baseline for calibrating the analysis results of the nonlinear analyzer, namely when the nonlinear analyzer ignores some redundant parameters due to strong correlation among parameters, if obvious linear or nonlinear correlation exists between the parameters and the tuning effect, the final sensitivity coefficient of the parameters can be improved by an aggregation mechanism, so that the occurrence of non-intuitive analysis results is avoided. Since the analysis result of the nonlinear analyzer is the main result, the mutual information value, the linear correlation degree, and the nonlinear correlation degree can be weighted and summed according to the ratio of 1:1:2 to obtain the sensitivity coefficient. The linear weighting mode can enable the finally obtained sensitivity coefficient to be more robust than a more extreme analysis result, and the sensitivity coefficient is less susceptible to extreme values.
Of course, the above-described polymerization method is merely exemplary, and the present embodiment is not limited thereto.
In summary, in this embodiment, multiple different machine learning algorithms may be aggregated to quantify complex influence relationships between multiple parameters in the target parameter set and tuning effects, provide a mechanism for screening sensitive parameters, optimize a tuning search space, and overall improve tuning efficiency. The nonlinear machine learning model can be responsible for capturing a complex nonlinear relationship between each of a plurality of parameters and the tuning effect, and further extracting and quantifying the nonlinear relationship through an interpretable algorithm. Meanwhile, a mutual information analysis module and a linear regression model are utilized to capture obvious mutual information and linear relations between a plurality of parameters and tuning effects, and based on the mutual information and the linear relations, quantitative results of the nonlinear relations are adjusted, so that the stability of overall identification of sensitive parameters is ensured.
It should be noted that, the execution subjects of each step of the method provided in the above embodiment may be the same device, or the method may also be executed by different devices. For example, the execution subject of steps 101 to 103 may be device a; for another example, the execution subject of steps 101 and 102 may be device a, and the execution subject of step 103 may be device B; etc.
In addition, in some of the flows described in the above embodiments and the drawings, a plurality of operations appearing in a specific order are included, but it should be clearly understood that the operations may be performed out of the order in which they appear herein or performed in parallel, the sequence numbers of the operations such as 101, 102, etc. are merely used to distinguish between the various operations, and the sequence numbers themselves do not represent any order of execution. In addition, the flows may include more or fewer operations, and the operations may be performed sequentially or in parallel.
Fig. 4 is a schematic structural diagram of a computing device according to another exemplary embodiment of the present application. As shown in fig. 4, the computing device includes: a memory 44 and a processor 41.
A processor 41 coupled to the memory 44 for executing a computer program in the memory 44 for:
obtaining a target parameter set to be optimized, wherein the target parameter set comprises a plurality of parameters;
performing simulation tuning test on the target parameter set to obtain a tuning effect test value;
according to the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values, determining the nonlinear correlation degree between each of the plurality of parameters and the tuning effect;
selecting sensitive parameters from the target parameter group according to the nonlinear correlation degree;
And performing parameter tuning on the object to be tuned according to the sensitive parameters.
In an alternative embodiment, the processor 41 is configured to, when determining the nonlinear correlation between each of the plurality of parameters and the tuning effect according to the parameter values of the target parameter set in the simulated tuning test process and the corresponding tuning effect test values:
Providing the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values for the machine learning model, so that the machine learning model captures the nonlinear relation between each of a plurality of parameters in the target parameter set and the tuning effect;
and quantifying the nonlinear relation captured by the machine learning model by using a model interpreter to generate nonlinear correlation between each of the plurality of parameters and the tuning effect.
In an alternative embodiment, processor 41 is configured to, when quantifying the nonlinear relationship captured by the machine learning model using the model interpreter to generate a nonlinear correlation between each of the plurality of parameters and the tuning effect:
constructing more parameter values and corresponding tuning effect test values in the simulated tuning test process according to the captured nonlinear relation by utilizing a machine learning model;
providing the parameter values generated by performing simulation tuning test operation on the target parameter set and constructed by the machine learning model and the corresponding tuning effect test values to a model interpreter;
and generating nonlinear correlations between each of the plurality of parameters and the tuning effect by using a model interpreter.
In an alternative embodiment, processor 41, when generating a non-linear correlation between each of the plurality of parameters and the tuning effect using the model interpreter, is configured to:
In a model interpreter, adopting a saprolite shape algorithm, and analyzing the contribution degree of each of a plurality of parameters to the tuning effect according to the parameter value and the corresponding tuning effect test value;
and determining the nonlinear correlation between each of the plurality of parameters and the tuning effect based on the contribution degree of each of the plurality of parameters to the tuning effect.
In an alternative embodiment, processor 41, when selecting the sensitive parameter from the set of target parameters based on the non-linear correlation, is configured to:
Obtaining linear relativity and/or mutual information values between each of the plurality of parameters and the tuning effect;
Correcting the nonlinear correlation according to the linear correlation and/or the mutual information value to obtain the sensitivity indexes corresponding to the parameters;
And selecting the sensitive parameters from the target parameter group according to the sensitive index.
In an alternative embodiment, processor 41, when obtaining the linear correlation between each of the plurality of parameters and the tuning effect, is configured to: performing linear analysis on the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values to generate linear correlation between each of the plurality of parameters and the tuning effect;
The processor 41 is configured to, when acquiring the mutual information values between each of the plurality of parameters and the tuning effect: and carrying out mutual information analysis on the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values so as to generate mutual information values between each of the plurality of parameters and the tuning effect.
In an alternative embodiment, when the processor 41 performs mutual information analysis on the parameter values of the target parameter set in the process of simulating the tuning test and the corresponding tuning effect test values to generate mutual information values between the multiple parameters and the tuning effects, the processor is configured to:
Aiming at the target tuning effect, at least one tuning effect test value corresponding to the target tuning effect and a parameter value in at least one target simulation tuning test process are selected from tuning effect test values generated in the simulation tuning test process;
performing mutual information analysis on the parameter values and the corresponding tuning effect test values in the at least one target simulation test process to generate mutual information values between a plurality of parameters and the target tuning effect;
The target tuning effect is any one of tuning effects.
In an alternative embodiment, processor 41 is configured to, when correcting the nonlinear correlation according to the linear correlation and the mutual information value to obtain the sensitivity indexes corresponding to the parameters respectively:
determining weights corresponding to the linear correlation, the mutual information value and the nonlinear correlation;
Under the target parameters, weighting linear correlation, mutual information value and nonlinear correlation according to the weights to obtain sensitivity coefficients corresponding to the target parameters;
wherein the target parameter is any one of a plurality of parameters.
In an alternative embodiment, processor 41 is configured to, when performing a simulated tuning test on the set of target parameters to obtain a tuning effect test value:
And performing simulation tuning test on the target parameter set by adopting a benchmark test benchmark technology so as to obtain a tuning effect test value.
Further, as shown in fig. 4, the computing device further includes: communication component 42, power component 43, and the like. Only some of the components are schematically shown in fig. 4, which does not mean that the computing device only includes the components shown in fig. 4.
It should be noted that, for the technical details of the embodiments of the computing device, reference may be made to the related descriptions of the embodiments of the method described above, which are not repeated herein for the sake of brevity, but should not cause any loss of protection scope of the present application.
Accordingly, embodiments of the present application also provide a computer-readable storage medium storing a computer program that, when executed, is capable of implementing the steps of the method embodiments described above that are executable by a computing device.
The memory of FIG. 4 described above is used to store a computer program and may be configured to store various other data to support operations on a computing platform. Examples of such data include instructions for any application or method operating on a computing platform, contact data, phonebook data, messages, pictures, videos, and the like. The memory may be implemented by any type of volatile or nonvolatile memory device or combination thereof, such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk.
The communication assembly of fig. 4 is configured to facilitate wired or wireless communication between the device in which the communication assembly is located and other devices. The device where the communication component is located can access a wireless network based on a communication standard, such as a mobile communication network of WiFi,2G, 3G, 4G/LTE, 5G, etc., or a combination thereof. In one exemplary embodiment, the communication component receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication component further comprises a Near Field Communication (NFC) module to facilitate short range communications. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, ultra Wideband (UWB) technology, bluetooth (BT) technology, and other technologies.
The power supply assembly shown in fig. 4 provides power for various components of the device in which the power supply assembly is located. The power components may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power for the devices in which the power components are located.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of computer-readable media.
Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.
The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and variations of the present application will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.
Claims (10)
1. A method of parameter tuning, comprising:
obtaining a target parameter set to be optimized, wherein the target parameter set comprises a plurality of parameters;
performing simulation tuning test on the target parameter set to obtain a tuning effect test value;
providing the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values for a machine learning model;
Quantifying a nonlinear relation between each of a plurality of parameters in the target parameter set captured by the machine learning model and the tuning effect by using a model interpreter so as to determine a nonlinear correlation between each of the plurality of parameters and the tuning effect;
selecting sensitive parameters from the target parameter group according to the nonlinear correlation degree;
And performing parameter tuning on the object to be tuned according to the sensitive parameters.
2. The method of claim 1, the quantifying the nonlinear relationship captured by the machine learning model with a model interpreter to produce a nonlinear correlation between each of the plurality of parameters and a tuning effect, comprising:
Constructing parameter values and corresponding tuning effect test values in a more simulated tuning test process according to the captured nonlinear relation by utilizing the machine learning model;
Providing the parameter values generated by the simulation tuning test operation on the target parameter set and constructed by the machine learning model and the corresponding tuning effect test values to the model interpreter;
Generating a nonlinear correlation between each of the plurality of parameters and the tuning effect by using the model interpreter.
3. The method of claim 2, the generating, with the model interpreter, a nonlinear correlation between each of the plurality of parameters and a tuning effect, comprising:
In the model interpreter, a saprolite shape algorithm is adopted, and the contribution degree of each of the multiple parameters to the tuning effect is analyzed according to the parameter value and the corresponding tuning effect test value;
And determining the nonlinear correlation between each of the plurality of parameters and the tuning effect based on the contribution degree of each of the plurality of parameters to the tuning effect.
4. The method of claim 1, the selecting a sensitive parameter from the set of target parameters according to the non-linear correlation, comprising:
acquiring linear relativity and/or mutual information values between each of the parameters and the tuning effect;
correcting the nonlinear correlation according to the linear correlation and/or the mutual information value to obtain sensitivity indexes corresponding to the parameters;
And selecting sensitive parameters from the target parameter group according to the sensitive index.
5. The method of claim 4, the obtaining a linear correlation between each of the plurality of parameters and a tuning effect, comprising: performing linear analysis on the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values to generate linear correlation between each of the plurality of parameters and the tuning effect;
The obtaining the mutual information value between each of the plurality of parameters and the tuning effect includes: and carrying out mutual information analysis on the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values so as to generate mutual information values between the multiple parameters and the tuning effect.
6. The method of claim 5, wherein performing mutual information analysis on the parameter values of the target parameter set in the simulated tuning test process and the corresponding tuning effect test values to generate mutual information values between each of the plurality of parameters and the tuning effect comprises:
Aiming at a target tuning effect, selecting at least one tuning effect test value corresponding to the target tuning effect from tuning effect test values generated in the simulation tuning test process and taking the value of a parameter in the at least one target simulation tuning test process;
performing mutual information analysis on the parameter values and the corresponding tuning effect test values in the at least one target simulation test process to generate mutual information values between the plurality of parameters and the target tuning effect;
Wherein the target tuning effect is any one of the tuning effects.
7. The method of claim 4, wherein correcting the nonlinear correlation according to the linear correlation and the mutual information value to obtain the sensitivity indexes corresponding to the parameters respectively comprises:
determining weights corresponding to the linear correlation degree, the mutual information value and the nonlinear correlation degree respectively;
Under a target parameter, carrying out weighted summation on the linear correlation, the mutual information value and the nonlinear correlation according to the weight so as to obtain a sensitivity coefficient corresponding to the target parameter;
wherein the target parameter is any one of the plurality of parameters.
8. The method of claim 1, the performing an analog tuning test on the set of target parameters to obtain a tuning effect test value, comprising:
And performing simulation tuning test on the target parameter set by adopting a benchmark test benchmark technology so as to obtain a tuning effect test value.
9. A computing device comprising a memory and a processor;
The memory is used for storing one or more computer instructions;
the processor is coupled to the memory for executing the one or more computer instructions for:
obtaining a target parameter set to be optimized, wherein the target parameter set comprises a plurality of parameters;
performing simulation tuning test on the target parameter set to obtain a tuning effect test value;
providing the parameter values of the target parameter set in the simulation tuning test process and the corresponding tuning effect test values for a machine learning model;
Quantifying a nonlinear relation between each of a plurality of parameters in the target parameter set captured by the machine learning model and the tuning effect by using a model interpreter so as to determine a nonlinear correlation between each of the plurality of parameters and the tuning effect;
selecting sensitive parameters from the target parameter group according to the nonlinear correlation degree;
And performing parameter tuning on the object to be tuned according to the sensitive parameters.
10. A computer-readable storage medium storing computer instructions that, when executed by one or more processors, cause the one or more processors to perform the parameter tuning method of any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111404891.3A CN114168216B (en) | 2021-11-24 | 2021-11-24 | Parameter tuning method, device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111404891.3A CN114168216B (en) | 2021-11-24 | 2021-11-24 | Parameter tuning method, device and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114168216A CN114168216A (en) | 2022-03-11 |
| CN114168216B true CN114168216B (en) | 2024-04-26 |
Family
ID=80480346
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111404891.3A Active CN114168216B (en) | 2021-11-24 | 2021-11-24 | Parameter tuning method, device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114168216B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1987905A (en) * | 2006-12-14 | 2007-06-27 | 东华大学 | Nerve network input parameter screening method based on fuzzy logic |
| CN108959741A (en) * | 2018-06-20 | 2018-12-07 | 天津大学 | A kind of parameter optimization method based on marine physics ecologic coupling model |
| CN110405343A (en) * | 2019-08-15 | 2019-11-05 | 山东大学 | A kind of laser welding process parameter optimization method of the prediction model integrated based on Bagging and particle swarm optimization algorithm |
| CN110825629A (en) * | 2019-10-31 | 2020-02-21 | 深圳市商汤科技有限公司 | Parameter tuning method and related device |
| CN111025899A (en) * | 2019-11-21 | 2020-04-17 | 复旦大学 | Nonlinear dynamic quality system prediction method |
| CN111651887A (en) * | 2020-06-03 | 2020-09-11 | 中国科学院华南植物园 | Method for analyzing uncertainty of parameters of numerical model |
| CN111859799A (en) * | 2020-07-14 | 2020-10-30 | 西安交通大学 | Method and device for evaluating data accuracy based on complex electromechanical system coupling relation model |
| CN112015620A (en) * | 2020-08-19 | 2020-12-01 | 浙江无极互联科技有限公司 | Method for automatically adjusting and optimizing parameters of website service end system |
| CN112015619A (en) * | 2020-08-19 | 2020-12-01 | 浙江无极互联科技有限公司 | Method for optimizing and screening core key indexes of system through parameters |
| CN112463763A (en) * | 2020-11-19 | 2021-03-09 | 东北大学 | RF algorithm-based MySQL database parameter screening method |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI549007B (en) * | 2013-02-07 | 2016-09-11 | 先知科技股份有限公司 | Method for searching and analyzing process parameters and computer program product thereof |
| TWI625682B (en) * | 2017-12-01 | 2018-06-01 | 財團法人工業技術研究院 | Methods, systems and non-transitory computer-readable medium for parameter optimization |
| CN110162857A (en) * | 2019-05-14 | 2019-08-23 | 北京工业大学 | A kind of flexible measurement method for surveying parameter towards complex industrial process difficulty |
-
2021
- 2021-11-24 CN CN202111404891.3A patent/CN114168216B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1987905A (en) * | 2006-12-14 | 2007-06-27 | 东华大学 | Nerve network input parameter screening method based on fuzzy logic |
| CN108959741A (en) * | 2018-06-20 | 2018-12-07 | 天津大学 | A kind of parameter optimization method based on marine physics ecologic coupling model |
| CN110405343A (en) * | 2019-08-15 | 2019-11-05 | 山东大学 | A kind of laser welding process parameter optimization method of the prediction model integrated based on Bagging and particle swarm optimization algorithm |
| CN110825629A (en) * | 2019-10-31 | 2020-02-21 | 深圳市商汤科技有限公司 | Parameter tuning method and related device |
| CN111025899A (en) * | 2019-11-21 | 2020-04-17 | 复旦大学 | Nonlinear dynamic quality system prediction method |
| CN111651887A (en) * | 2020-06-03 | 2020-09-11 | 中国科学院华南植物园 | Method for analyzing uncertainty of parameters of numerical model |
| CN111859799A (en) * | 2020-07-14 | 2020-10-30 | 西安交通大学 | Method and device for evaluating data accuracy based on complex electromechanical system coupling relation model |
| CN112015620A (en) * | 2020-08-19 | 2020-12-01 | 浙江无极互联科技有限公司 | Method for automatically adjusting and optimizing parameters of website service end system |
| CN112015619A (en) * | 2020-08-19 | 2020-12-01 | 浙江无极互联科技有限公司 | Method for optimizing and screening core key indexes of system through parameters |
| CN112463763A (en) * | 2020-11-19 | 2021-03-09 | 东北大学 | RF algorithm-based MySQL database parameter screening method |
Non-Patent Citations (1)
| Title |
|---|
| 一种基于敏感性分析的改进参数筛选方法;黄欣;李倩玉;骆亦其;薛巍;;电子技术应用(第12期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114168216A (en) | 2022-03-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110163261B (en) | Unbalanced data classification model training method, device, equipment and storage medium | |
| KR101967089B1 (en) | Convergence Neural Network based complete reference image quality evaluation | |
| CN117435505B (en) | Visual generation method of performance test script | |
| CN115757745A (en) | Service scene control method and system based on artificial intelligence and cloud platform | |
| CN114168216B (en) | Parameter tuning method, device and storage medium | |
| Bhorkar et al. | DeepAuto: A hierarchical deep learning framework for real-time prediction in cellular networks | |
| CN114490413A (en) | Test data preparation method and device, storage medium and electronic equipment | |
| CN115730507A (en) | Model engine construction method, kernel function processing method, device and storage medium | |
| CN118097355A (en) | Alarm information processing method, equipment and medium based on ensemble learning | |
| CN112906883A (en) | Hybrid precision quantization strategy determination method and system for deep neural network | |
| CN117113061A (en) | Cross-receiver radiation source fingerprint identification method and device | |
| CN118052310A (en) | Time sequence prediction optimization method, device and storage medium | |
| CN113656279B (en) | Code odor detection method based on residual network and metric attention mechanism | |
| CN109086201A (en) | Automatic software test method and system | |
| CN115358410A (en) | Method, device and equipment for enhancing field of pre-training model and storage medium | |
| CN110895508B (en) | Method and device for generating traversal test path | |
| CN117973904B (en) | Intelligent manufacturing capacity analysis method and system | |
| CN116665020B (en) | Image recognition method, device, equipment and storage medium based on operator fusion | |
| CN117971705B (en) | Intelligent interface automatic test system and method based on customized flow insight | |
| CN118522340B (en) | Method, device, equipment and storage medium for monitoring biodiversity of regional scope | |
| CN118074806B (en) | Optical amplifier gain adjusting method and equipment based on machine learning | |
| CN116935102B (en) | Lightweight model training method, device, equipment and medium | |
| CN116527411B (en) | Data security intelligent protection model construction method and device and collaboration platform | |
| US20230306377A1 (en) | Offline machine learning for automatic action determination | |
| CN117172789A (en) | Risk assessment model construction method and device for suspicious transaction monitoring |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |